April, ’24] 
HARTZELL: INTERPRETATION OF CODLING MOTH DATA 
191 
before it would have been possible to definitely conclude that, for 
that season , larger amounts of arsenate of lead gave better control. 
Table 5. Means and Probable Errors Computed on the Basis of Sixteen 
Trees per Plat 
Plat 
Mean 
Plats 
Compared 
Difference 
Difference 
divided by 
probable error 
Total Infested Apples 
Check 
993.00 ±27.49 
Check & 1 
489.25 ±38.53 
12.70 
1 
503.75 ±27.00 
Check & 2 
230.00 ±55.15 
4.17 
2 
763.00 ±47.81 
Check & 3 
611.71 ±32.27 
18.96 
3 
381.29±16.91 
1 and 2 
259.25 ±54.91 
4.72 
1 and 3 
122.46 ±31.86 
3.84 
2 and 3 
381.71 ±50.71 
7.53 
Worthless Apples 
Check 
489.00±18.21 
Check & 1 
290.00 ±19.84 
14.62 
1 
199.00 ± 7.88 
Check & 2 
331.00 ±21.40 
15.47 
2 
1.58.00 ±11.24 
Check & 3 
386.14±19.17 
20.14 
3 
102.86 ± 5.98 
1 and 2 
41.00 ±13.73 
2.99 
1 and 3 
96.14 ± 9.89 
9.72 
2 and 3 
55.14±12.73 
4.33 
Since a single set of field tests constitutes one observation in a universe 
of experiments, it is obvious that recommendations to the farming 
public should not be made until the tests have been repeated during 
the same and different seasons. Not less than ten such trials should 
be made and 16 would be better before making changes in the spray 
schedule. To increase the amount of poison in sprays means an ad¬ 
ditional cost of thousands of dollars to fruit growers, therefore, it be¬ 
hooves the experimenter to have abundant data that has been care¬ 
fully analyzed and found significant before changes should be advocated. 
Useful Probable Error Formulae for Entomologists 
The economic entomologist will have occasion to use several formulae 
for probable errors which are not generally found in texts on statistical 
methods. These have been included for convenience. 
If a general mean is desired of several quantities affected with the 
errors ei, e 2 , etc., the formula is E gm = ± 
+ e 2 + . . .e n , 
n 
in which n is the number of values used in computing the general mean. 
